Product data quality sounds like polish and is actually plumbing: every marketplace renders your catalog through its own templates, search engines, and shipping calculators, and each one reads specific fields. A blank weight is a wrong shipping quote on three channels; a missing category is a listing rejection; a thin title is invisible search placement everywhere at once. The bar is completeness in the SOURCE, because multichannel copies whatever you have, at machine speed.
The six field groups, by consequence
1. Identity. SKU per variant, GTIN where required, brand, and model/MPN where applicable. Consequences of gaps: failed matching, catalog mismatches, listing rejections in GTIN-required categories.
2. Physical attributes. Weight and dimensions (unit and packaged), material, and whatever the category legally or practically demands. Consequences: shipping calculators quote fiction, rate strategies break, oversize surcharges arrive as surprises, and fee previews mislead, FBA fees are dimension-driven.
3. Categorization. The right category per channel plus the item specifics/attributes that category expects. Consequences: the largest single lever on marketplace search visibility, and the most common listing-push failure when tools submit to category-strict platforms, category mapping is its own discipline.
4. Imagery. Platform-compliant primary images (background rules, minimum resolution), per-variant images (the picker must change the photo), and assets hosted where YOU control them, the inkFrog lesson.
5. Pricing inputs. Base price, cost (for margin truth), and MAP floors where they exist, the raw material per-channel rules compute from. Consequence of gaps: rules that cannot protect margins they cannot see.
6. Compliance and trust. Condition, country of origin, safety and battery flags where categories demand them, and honest description text. Consequences range from listing blocks to the expensive kind of buyer surprise, returns with reason codes attached.
The afternoon audit
Export the catalog and count blanks per field group, the readiness-test export doubles as the audit surface:
- Sort by blank density: fields empty on >10 percent of SKUs are systematic, not accidental, fix the intake process, then the rows.
- Fix by consequence order: identity first (it blocks everything), then physical attributes (money leaks quietly), then categories, images, pricing inputs, compliance.
- Spot-check honesty, not just presence: a filled-but-wrong weight is worse than a blank one, it fails silently. Ten random SKUs against a tape measure and a scale calibrates your trust.
- Assign the field owner: every future product enters through a checklist that fills the six groups at creation, data quality is an intake property; audits just catch drift.
Where the data lives and flows
One source of truth (your store), attributes complete there, flowing to channels through the connection, never re-typed per channel. Channel-specific adaptations (title styles, category mappings) are transformations ON the source data, not forks of it, forks are how the Amazon listing gets the corrected weight and eBay keeps quoting the old one.
Common questions
How complete is complete enough?
Every field a connected channel reads, filled and true. The practical floor: identity + weight/dimensions + category + one compliant image per variant. Below that, listings fail; above it, quality is conversion work.
Do descriptions need to differ per channel?
Titles benefit from per-channel style; descriptions can share a truthful core with channel formatting. What must NOT differ: the facts.
Who should own product data in a small team?
One named owner for the intake checklist, whoever creates products fills the six groups. Shared ownership of data quality means nobody owns the blanks.
Can software fix data quality?
Tools surface gaps (field audits, failed-push reasons) and stop the bleeding (one source, no re-typing). The facts themselves, weights, materials, honest photos, come from the person holding the product.
Complete once, correct everywhere
One source of truth flowing to every channel, with the sync and rules that keep it honest, from $49/month with unlimited orders. See pricing.